Annals of the Rheumatic Diseases, 1983, 42, 75-81

Morphological characteristics of monosodium urate:a transmission electron microscopic study of intactnatural and synthetic crystalsHERNANDO PAUL, ANTONIO J. REGINATO, ANDH. RALPH SCHUMACHER

From the University of Pennsylvania School of Medicine and Veterans Administration Medical Center,Philadelphia, PA 19104, USA

SUMMARY Transmission electron microscopic studies of synthetic and natural monosodium uratecrystals dried on formvar coated grids showed identical internal structures in all crystals. At highermagnification the crystals' surface showed angular or wavy irregularities, and more rarely some

crystals appeared to have other tiny crystals on the surface. Protein-like surface coating was notobserved except in crystals from one asymptomatic patient in whom synovial fluid was loaded withmonosodium urate crystals, but no inflammatory cells were present. Heated synthetic monosodiumurate crystals retained the ultrastructural characteristics in their interior but they lost their needleor rod-like shape. Transmission electron microscopic study of monosodium urate crystals dried onformvar coated grids provides a quick method of investigating crystal ultrastructure.

Monosodium urate (MSU) crystal structure as seenby compensated polarised light microscopy has beendescribed in detail,' but routine transmission electronmicroscope (TEM) studies of thin sections of thesynovial membrane and tophi have failed to show theultrastructure of these crystals.2 3 Since MSU crystalsare dissolved out during tissue processing for TEM,onlv crvstal outlines have been observed. Recentstudies of crystal surface-protein interactions,3 4adsorption of lysosomal enzymes to crystals,5 andnucleation of MSU crystals' suggest that furtherknowledge of the ultrastructure of MSU might behelpful in understanding the pathogenesis of gout.However, only very few TEM studies have beenundertaken on intact MSU crystals. In 1967 Riddle etal. briefly described the electron microscope appear-ance of such crvstals and the alteration of the internalstructure by electron beam exposure.7 More recentlyPritzker et al., using scanning and TEM techniques,reported that the internal architecture of urate wasdifferent in synovial fluid (SF) and tophi. The ultra-structure of crystals in tophi suggested the presenceof nucleating material.8

Accepted for publication 22 January 1982.Correspondence to Professor A. J. Reginato MD. Veterans Admini-stration Medical Center, University and Woodland Avenues,Philadelphia, PA 19104, USA.

In this report we describe the ultrastructure ofheated and unheated intact synthetic MSU crystalsand our findings on natural MSU from SF, bursae,and tophi. The external appearance of syntheticcrystals was altered by heat. A similar internalarchitecture was found in synthetic crystals andnatural crystals from SF, bursae, and tophi. MostMSU crystals had smooth surfaces with evidence ofthick superficial coating on intact crystals from onlyone synovial fluid.

Materials and methods

MSU crystals were obtained from 15 synovial fluids,3 olecranon bursal fluids, one tophus of the helix ofthe ear, and one tophus adjacent to a knee joint. Tensynovial fluid specimens were aspirated from acutelyinflamed and 4 from chronically inflamed joints. OneSF was from an asymptomatic joint of a patient with amassive amount of MSU crystals in the synovialmembrane and synovial fluid but without an acuteinflammatory reaction.

In order to compare the ultrastructure of naturalMSU crystals with synthetic crystals, samples ofsynthetic MSU were prepared in water by themethods of Denko and Whitehouse9 and of McCartyand Faires."0 Crystals were examined after beingheated at 180°C for 3 hours and unheated.

75

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

76 Paul, Reginato, Schumacher

One drop aspirated from the samples containingMSU crystals was promptly placed on a clean glassslide. Six formvar-coated EM grids were floated onthe drop for 10 seconds. To obtain samples withthickness suitable for TEM studies excessive materialon the grids was removed gently with the use of filterpaper. The grids were then air dried and observedwith compensated polarised light microscopy. Thosecontaining negatively birefringent crystals wereexamined unstained in a Zeiss EM 10 TEM with a 60kV beam. The morphology of MSU crystals wasstudied in over 200 photomicrographs.Four samples of natural MSU crystals and one of

heated synthetic crystals were studied by x-ray dif-fraction. A Debbye-Scherrer powder camera (NorthAmerican Phillips) of 114 6 mm diameter withchromium K alpha radiation and a vanadium filterwas used. Measurements were based on wave length2-29092A (22.9092 nm). The specimens wereexposed for 20 hours at 60 kV and 20 mA. All 5specimens studied showed exclusively MSU crystals.

Results

NATURAL MSU CRYSTALSUrate crystals including a range varying in size,shape, and ultrastructure were seen in each indi-vidual sample. Birefringence by polarised light wasstrong or weakly negative. Similar general findingswere observed in SF, bursae, and tophi (Fig. 1).Although variation in size had been appreciated bypolarised light microscopy, the range of differentshapes and extremes of small size had not beenreadily detectable.

/,

Fig. 1 Ear tophus. MSU crystals exhibiting variable shape,size, and electron density. These are similar to those MSUobserved in synovial fluid and articular tophi. Stacking ofcrystals is seen at the arrows. ( x2860).

Sizes varied considerably, but there was a tendencyfor the crystals to be smaller in SF. Crystal length ingeneral ranged from 1 to 40,m in bursae and tophi,while it was between 1 and 20 ,um in SFs. Mostcrystals were long with parallel sides and 2 blunt ends(Fig. 2A). Other crystals were long with parallel sidesand one somewhat less blunt end, though ends werenot truly needle-like (Fig. 2B). A few crystals werevery short (0.5 ,um) and rod-shaped with 2 bluntends. A common finding was stacking together of 2 ormore crystals (Figs. 1, 2B). Some of these twinned orstacked crystals appeared to have structural con-tinuity between individual crystals. In other stacked

Fig. 2 Different shapes and sizes ofMSU crystals. 2A. Most MSU crystals were long with parallel sides and blunt ends.Olecranon bursal fluid ( x 12 900). 2B. Other MSU crystals exhibiting parallel sides with one blunt end and others less bluntbut not truly needle-like. Apparent stacking ofcrystal components is also seen, with continuity ofcrvstal structure in the centre.Synovial fluid ( x 10 300).

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

Morphological characteristics ofmonosodium urate 77

forms this appearance seemed to arise from super-imposition of crystals.The crystal surface was most often smooth at lower

magnifications (Figs. 1, 2), but at higher magnifica-tions angular or curved irregularities of the crystalsurface were observed in most SF, bursae, and tophi(Figs. 3A, 3B). Some crystals appeared to have othertiny crystals on the surface (Fig. 3C). A few of thosetiny crystals exhibited apparent continuity with thelarger crystal. Others had different ultrastructure andwere actually morphologically reminiscent of calciumpyrophosphate dihydrate (CPPD) because of theirgreater density with round and more uniformlyarranged holes.'" The latter ultrastructurally dif-ferent crystals were observed in bursae (Fig. 3C) andSF but not in the 2 tophi examined.

Diligent examination of the surface failed to revealany evidence of definite protein-like coating on mostcrystals. A thick fluffy coating was observed, how-ever, on MSU crystals in the SF from the one patientwho was asymptomatic (and without SF inflammat-ory cells) (Fig. 3D). Many crystals had a sharply

demarcated, less dense homogeneous surface layerthat might possibly be a coating or more likely ismerely one appearance of the surface crystalmaterial. There was partial disappearance of thefluffy coating after treatment with hyaluronidase, butnot of the homogeneous material on other crystals.The internal architecture as seen by TEM also

varied considerably. MSU crystals showing differentgrades of electron density and variable internalarchitecture were observed in all specimens of SF,bursae, or tophi (Figs. 2A, 4C). The interior of thecrystals developed an electron-lucent network whichwas artefactually induced by exposure to the electronbeam. At the very beginning of electron beam ex-posure the crystals appeared uniformly electron-dense (Fig. 4A). Within a few seconds many smallelectron-lucent areas were observed (Fig. 4B). Somecrystals appeared to become stabilised and showedno further changes, even after prolonged beamexposure (Figs. 2A, 3C, 3D). This was usually true ofthe smaller and/or thinner crystals. In larger andthicker crystals the electron lucencies enlarged while

3A

3 C 30

Fig. 3 Surface characteristic and coating ofMSU crystals. 3A. MSU crystal showing less dense homogeneous surface withangular irregularities of the surface. Synovial fluid. ( x1 7 950). 3B. MSU crystal with wavy irregularities on the surface.Synovial fluid. ( x2 780). 3C. MSU crystal with tiny crystals on its surface. Some satellite crystals exhibit internal structure likeMSU crystals and showed structural continuity with the large crystals (arrow), while others appear to be only attached to thesurface (double arrow). Olecranol bursal fluid ( x 1 7 950). 3D. Abundant granular coating on MSU crystal in the synovialfluid ofan asvmptomatic gout)? patient with massive amounts ofurate crystals in the synovial fluid without inflammatory cells(x36 800).

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

78 Paul, Reginato, Schumacher

4'A

4

Fig. 4 Changes on MSU crystal structure by electron beam exposure. 4A. At the beginning, crystals appear uniformlyelectron-dense. Ear tophus. ( x15 900). 4B. After a few seconds a lucent network has developed. Ear tophus ( xiS 900).4C. Rarely apparent MSU crystals have larger ovoid electron-lucent spaces. Olecranon bursal fluid. (x99 400).

the dense areas appeared to aggregate into increas-ingly dense round or oblong structures, producing theappearance of an 'electron lucent tubular networkbordered by an electron dense shell'8 (Figs. 3A, 3B).Normally these structural changes were observed inthe central areas of the crystal, while the crystal endsdeveloped only smaller electron lucencies. In bursaevery occasional crystals showed larger ovoidelectron-lucent spaces (Fig. 4C).

SYNTHETIC MSU CRYSTALSUnder polarised light crystals synthesised over 72hours by the methods previously mentioned were

predominantly needle or rod shaped, were largerthan natural MSU, and always very strongly showednegative birefringence. By TEM synthetic crystalswere between 2 and 60 um and had parallel sideswith smooth surfaces (Fig. 5A). They had no tinycrystals on the surface and no fluffy coating wasobserved. They showed variable density anddeveloped structural changes induced by the electronbeam exposure similar to those of natural MSUcrystals. Heated MSU crystals retained their ultra-structural characteristics in their interior, but theylost their needle- or rod-like shapes, and the surfacebecame very irregular (Fig. 5B). A homogeneous

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

Morphological characteristics ofmonosodium urate 79

5A 5A

Fig. S Synthetic MSUcrystals. 5A. Synthetic MSUcrystalsarelargerthan naturalcrystals. Atlowmagnification theyexhibitsmooth surfaces and needle-like or rod shapes. Electron micrograph taken after minimal exposure to electron beam. ( x33 70).SB. Heated synthetic MSUcrystals retain the internal ultrastructural characteristics, but they acquire a bizarre appearance andirregular surface and lose their well defined rod- or needle-like shapes. The homogeneous surf(ace layer is seen at the arrow.

lighter surface structure, such as was seen on somenatural crystals, occurred on some of the syntheticcrystals with or without heating (Fig. 5B).

Discussion

The present study provides further information onthe morphological characteristics of natural andsynthetic MSU crystals. These crystals and otherwater soluble crystals such as steroid crystals'2 andcholesterol crystals"3 are usually lost during tissueprocessing, and only an electron-dense outline of thecrystal remains. CPPD and calcium oxalate'4 are notdissolved, but sections for TEM might cut throughcrystals making shape an imprecise criterion for theiridentification. By using the formvar-coated gridtechnique, crystal dissolution is avoided and crystalsretain their actual shapes. CPPD crystals continuebeing electron-dense and diffusely foamy with round,more uniformly distributed holes,'5 and are thus simi-lar in fine structure to calcium oxalate.'4 Cholesteroland steroid crystals appear uniformly electron-densewith almost no changes induced by the beam ex-posure. Thus all these crystals have their own charac-teristics, which are different from those of MSUcrystals. The occasional atypical crystals seen in bur-sae (Figs. 3C, 5C) are still unexplained and mightrepresent different morphology of MSU or minuteamount of some other crystal missed by x-ray diffrac-tion analysis.MSU crystals in our study showe.d some variability

in shape and internal architecture. The same range of

findings was consistently observed in SF, bursae,tophi, and synthetic crystals, indicating that MSUcrystals formed or found under different conditionshad a distinct ultrastructural resemblance. Previousobservations had suggested that urates precipitatedin tophi and SF appeared different.8 We were able toobserve similar crystals to those described by theseauthors, but they were seen in specimens from SF,bursae, or tophi. Since only few tophi were studiedby Pritzker et al.8 and by us, the differences mightrepresent sampling of crystals predominantly at dif-ferent stages of crystal growth or variations in indi-vidual patients, as well as different techniques ofprocessing crystals. In the present study the MStTcrystal size in SF was slightly smaller than in tophi orbursae. Smaller crystals have also been observed inMSU synthesised in SF than in water.'6 Factors suchas lysis of urate crystals by neutrophil peroxidases'7or other factors increasing MSU solubility'8 may alsobe responsible for the smaller size of MSU crystals inSF.The different grades of electron density and

internal architecture observed in our study appearedto result from the variability of crystal size and shape.Thinner crystals seemed to respond differently to theelectron beam. The variable ultrastructural networkshown here was clearly an artefact induced by theelectron beam exposure. We have, however, demon-strated in other studies that nonartefactual holes areidentifiable during the early growth of MSU crys-tals.'6 Whether these holes during growth, and anynucleating material that may be trapped in the

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

80 Paul, Reginato, Schumacher

crystal, influence the later changes seen in the elec-tron beam is not known.An interesting observation was the failure to show

any definite or significant surface coating on mostcrystals. The ability of the MSU crystal surface tointeract with proteins has been proposed,"9-2" andactual immunoglobulin coating on crystals in vivo hasbeen previously suggested.34 But morphologicalproof of location of immunoglobulin on the crystalsurface had not been presented. It is possible that theformvar-coated grid techniques used in our study didnot allow visualisation of small amounts of adherentmaterial on the surface of crystals, or that digestion ofmuch of the protein coat had already occurred by thelysosomal enzymes in the inflammatory fluids.22Pritzker et al.,8 using similar techniques to ours,observed amorphous substances coating the surfaceof MSU crystals in SF and bursae. Since their sampleswere preserved in ethanol, these authors interpretedtheir findings as precipitated 'mucin'. It is intriguingthat they observed smooth surfaced crystals on thejoint surfaces of asymptomatic gouty patients, whilein our study, on the contrary, the only SF in which adefinite crystal surface coat was observed came froman asymptomatic gouty patient. What this fluffy coat-ing contains and whether it is what preventedinflammation remains to be determined. Whetherthe less dense homogeneous material on the surfaceof many crystals is only part of the crystal wall or mayinclude coating material will be studied in the futurewith immuno EM techniques.Some very small crystals were also observed in

fluids from our patients. Many were on the surface ofother crystals; we do not know the exact nature ofthese crystals. X-ray diffration in one fluid containingsuch crystals showed only MSU, but very small cryst-als may not allow electron diffraction or give offsufficient x-rays for elemental analysis. Thus it ispossible that they represent urate or other uniden-tified crystals, such as even CPPD or other calciumsalts.The study of synthetic MSU crystals clearly

showed that heating may induce morphologicalchanges, especially on the crystal surface, destroyingthe well-defined needle-like or rod shapes. Neverthe-less, the internal architecture of heated crystalsappeared similar to unheated crystals. Mandel, usingx-ray diffraction and thermal gravimetric analysis,has postulated that heating of MSU crystals induceschanges in the chemical composition and crystalstructure .

The present studies demonstrate that drying verysmall amounts of fluids on formvar-coatcd gridsallows visualisation of whole MSU crystal ultrastruc-ture. Some of our observations confirm the findingspreviously reported by Pritzker and his co-workers,8

including the development of an internal tubularnetwork that can be seen after exposure to the elec-tron beam. However, in contrast to them we foundvirtually the same ultrastructure of crystals fromtophi, bursae, and SF. We have shown, so far, adistinctive fluffy coating on the surface of crystalsfrom one joint that had never been clinicallyinflamed. Whether the ultrastructural characteristicsmay help clarify factors involved in the inflammatoryproperties of MSU crystals remains uncertain. How-ever, since MSU crystals appear to have a charac-teristic ultrastructure, rapid examination on SF onformvar-coated grids by TEM, when available, offersanother aid to diagnosis in gouty arthritis, when nocrystals, or only unusual or very small crystals, areseen by polarised light.24 Scanning electron micros-copy has similarly been used to identify crystals insynovial fluid.25 Combinations of these newerapproaches offer the opportunity of many newadvances in the understanding of the role of crystalsor other particulate matter in arthritis.2628

The authors acknowledge the superb technical assistance of GildaClayburne, Susan Rothfuss, Marie Sieck. and the excellent secre-tarial help of Joanne Loguidice.The work was supported in part by grants from the McCabe

Foundation, Barsumian Fund, Veterans Administration, and fromCentro Nacional de Enfermedades Reumaticas MSAS UniversidadCentral de Venezuela and Fundacion Gran Mariscal de Ayacucho,Venezuela.

References

1 Phelps P, Steele A D. McCarty D J. Compensated polarized lightmicroscopy. Identification of crystals in synovial fluids from goutand psuedogout. JAMA 1968; 203: 508-12.

2 Agudelo C A, Schumacher H R. The synovitis of acute goutyarthritis. A light and electron microscopic study. Hum Pathol1973; 4: 265-79.

3 Hasselbacher P, Schumacher H R. Immunoglobulin in tophi andon the surface of monosodium urate crystals. Arthritis Rheum1978; 21: 353-61.

4 Kozin F, McCarty D J. Protein adsorption to monosodium urate,calcium pyrophosphate dihydrate and silica crystals. ArthritisRheum 1976; 19: 433-8.

5 Ginsberg M H, Kozin F, Chow D, May J, Skosey J L. Adsorptionof polymorphonuclear leukocyte lysosomal enzymes to mono-sodium urate crystals. Arthritis Rheum 1977; 20: 1538-42.

6 Tak H, Cooper S M, Wilcox W R. Studies on the nucleation ofmonosodium urate at 37C. Arthritis Rheum 1980; 23: 574-80.

7 Riddle J M, Bluhm G B, Barnhart M I. Ultrastructural study ofleukocytes and tirates in gouty arthritis. Ann Rheum Dis 1967;26: 389-401.

8 Pritzker K P H, Zahn C E, Nyburg S C, Luk S C, Houpt J B. Theultrastructure of urate crystals in gout. J Rheumatol 1978; 5:7-18.

9 Denko C W, Whitehouse M W. Experimental inflammationinduced by naturally occurring microcrystalline calcium salts.J Rheumatol 1976; 3: 54-62.

10 McCarty D J, Faires J S. A comparison of the duration of localanti-inflammatory effect of several adrenocorticosteroid esters.A bioassay technique. Curr Ther Res 1963; 5: 284-90.

11 Schumacher H R. Ultrastructural findings in chondrocalcinosisand pseudogout. Arthritis Rheum 1976; 19: 413-25.

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from

Morphological characteristics ofmonosodium urate 81

12 Gordon G V, Schumacher H R. Electron microscopic study ofdepot corticosteroid crystals with clinical studies after intra-articular injection. J Rheumatol 1979; 6: 7-14.

13 Parker F, Odland G F. Ultrastructural and lipid biochemicalcomparisons of human eruptive tuberous and planar xanthomas.Isr J Med Sci 1973; 9: 395-423.

14 Hoffman G S, Schumacher H R, Paul H, Cherian V, Ramsey A,Frank W A. Calcium oxalate crystals associated arthritis inchronic renal failure. Arthritis Rheum 1981; 24: S73 (abst).

15 Pritzker N P H, Phillips H, Luk S C, Koven I H, Kiss A,Houpt J B. Pseudotumor or temporomandibular joint destruc-tive calcium pyrophosphate dihydrate arthropathy.J Rheumatol1976; 3: 70-81.

16 Paul H, Schumacher H R. Sequential ultrastructural changes inthe growth ofmonosodium urate crystals. Arthritis Rheum 1981;24: 573 (abst).

17 Howell R R, Seegmiller J E. Uricolysis by human leukocytes.Nature 1962; 196: 482-3.

18 Kippen I, Klinenberg J R, Weinberger A, Wilcox W R. Factorsaffecting urate solubility in vitro. Ann Rheum Dis 1974; 33:313-7.

19 Kellermeyer R W, Breckenridge R T. The inflammatory processin acute gouty arthritis. Activation of Hageman factor by sodiumurate crystals.J Lab Clin Med 1965; 65: 307-15.

20 Naff G B, Byers P H. Complement as a mediator of inflamma-tion in acute gouty arthritis. I. Studies on the reaction between

human serum complement and sodium urate crystals.J Lab ClinMed 1973; 81: 747-60.

21 Campion D S, Bluestone R, Klinenberg J R. Displacement byuricosuric agents of sodium urate bound to human serum albu-min. Biochem Pharmacol 1974; 23: 1653-7.

22 Wailingford W R, McCarty D J. Differential membranolyticeffects of microcrystalline sodium urate and calcium pyroph-sophate dihydrate. J Exp Med 1971; 133: 100-12.

23 Mandel N S. Structural changes in sodium urate crystals onheating. Arthritis Rheum 1980; 23: 772-6.

24- Honig S, Gorevic P, Hoffstein S, Weissmann G. Crystal deposi-tion disease. Am J Med 1977; 63: 161-4.

25 Faure G, Netter P, Malman B, Steinmetz J, Duheille J,Gaucher A. Scanning electron microscopic study of microcryst-als implicated in human rheumatic diseases. Scanning ElectronMicroscopy 1980; 3: 163-78.

26 Crocker P R, Dieppe P A, Tyler T, Chapman S K,Willoughby D A. The identification of particulate matter inbiological tissues and fluids. J Pathol 1976; 121: 37-40.

27 Crocker P R, Doyle D V, Levison D A. A practical method forthe identification of particulate and crystalline material inparaffin-embedded tissue specimens. J Pathol 1980; 131:165-73.

28 Ali S Y, Griffiths S. New type of calcium phosphate crystals inarthritis cartilage. Semin Arthritis Rheum 1981; 11: suppl 1:124-6.

copyright. on F

ebruary 21, 2022 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.42.1.75 on 1 F

ebruary 1983. Dow

nloaded from